Application and Comparison of Different Optimization Algorithms in Stand Management Optimization

An artificial bee colony algorithm(ABC) with few control parameters and strong optimization mechanism was introduced to optimize stand management measures based on the maximum net present value(NPV) , the parameters configuration and characteristics of ABC algorithm are also discussed by comparing with Hooke & Jeeves direct search algorithm, differential evolution algorithm(DE) , evolutionary strategy algorithm(ES) and particle swarm optimization algorithm(PSO) . Through simulating the growth and management processes of an initial Korean pine plantation(planting density 2500 trees/ha, site index = 16 m, age = 10 years) and optimizing its management schedule, the parameters of the ABC algorithm were systematically explored to determine their optimal configuration. The results showed that NPV increases with the increase of the swarm size

. When swarm size was 90, NPV consistently exceeded 385,500 yuan/ha. Comparative results for different algorithms revealed that, according to the mean NPV of repeated optimization runs, the ranking of the algorithms was PSO > ABC > DE > ES > HJ. According to the coefficient of variation, the ranking was DE(best) < PSO < ABC < ES < HJ. When the population size(swarm size) was reduced to 5 while the other parameters were kept in near-optimal values, the ranking of the population-based algorithms in terms of mean NPV was ABC > DE > PSO > ES. In terms of the coefficient of variation, the ranking was DE < ABC < PSO < ES. Overall, DE, PSO, and ABC algorithms exhibited good performance and robustness as well as the ability to maintain diversity among the population of the solutions The computing times were shorter for ABC, as compared to DE and PSO. This study systematically evaluates the performance of five forest management optimization algorithms under optimal parameter configurations. Overall, DE, PSO, and ABC algorithms exhibit outstanding performance with strong robustness and the ability to maintain diversity in candidate solutions. However, in handling complex optimization problems, the ABC algorithm demonstrates superior execution efficiency. By comparison, this study evaluated the feasibility of five algorithms to optimize management measures, and provided scientific support for the application of ABC algorithm in stand management optimization.